GitLab

The portable implementation in base is nicely written and tested but
could be made way faster by using SIMD and crypto instructions.

Enters BoringSSL: its SHA1 implementation is highly optimized for
many architectures, actively maintained and tested.

It will make SHA1 *way* faster on Intel (6x), ARMv7 (8x) and
ARMv8 (between 16x to 21x faster).

Plus adds a performance benchmark to measure the speed (MB/s) of
the hash implementation.

Unfortunately, thanks to NaCL we still got keep the portable
implementation around for the time being (until it is finally
removed).

Change-Id: Ifc4b6a82a0d36e5157b1fa6d94760c32197b570c
Bug: 974341
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1661110Reviewed-by: Adenilson Cavalcanti <cavalcantii@chromium.org>
Reviewed-by: Daniel Cheng <dcheng@chromium.org>
Reviewed-by: David Benjamin <davidben@chromium.org>
Commit-Queue: Adenilson Cavalcanti <cavalcantii@chromium.org>
Cr-Commit-Position: refs/heads/master@{#678434}